When Shifa Abbas sat down with Brian McCrindle on Subnet Magazine, the conversation felt less like an interview and more like a journey through the evolving world of decentralized intelligence.
Brian, one of the lead engineers behind Macrocosmos and Bitstarter, carries a rare mix of deep technical expertise and reflective curiosity.
Together, they explored everything from semiconductors and Swarm training to the philosophy of building in an open, unpredictable ecosystem like Bittensor.
From Physics Labs to Decentralized Intelligence
Shifa began by asking Brian about his background, and it quickly became clear that his path into Web3 was anything but typical.
With a degree in engineering physics and a master’s in electrical and computer engineering, he spent years studying semiconductor-based lasers and machine learning models that improve manufacturing yield.
He described how a one percent improvement in efficiency could save millions, and how that experience taught him the value of precision, patience, and iteration; the principles that now guide his engineering work on decentralized systems.
Discovering Bittensor
When Shifa turned the conversation toward Bittensor, Brian shared how his introduction to the network happened almost by accident.
After leaving his job, he reconnected with Steffan Cruz who at the time was serving as CTO of the Opentensor Foundation. Over drinks by a campfire in Mexico, Stephan told him about a protocol that could decentralize intelligence itself.
“I didn’t know anything about crypto,” Brian recalled with a laugh. “But it made sense. The incentives were elegant. I just wanted to be part of it.”
That night marked the beginning of his journey into the Bittensor ecosystem, joining the foundation during its earliest, most experimental phase. A time defined by both chaos and possibility.
The Birth of Macrocosmos
As Bittensor matured and the Opentensor Foundation moved toward dissolution to achieve decentralization, Brian and a small group of collaborators formed Macrocosmos, a team that continues to run several subnets including Apex (Subnet 1), IOTA (Subnet 9), and Data Universe (Subnet 13).
Under Shifa’s guidance, Brian described how Macrocosmos built itself around a culture of curiosity and ownership. “We’ve grown into a team of about thirty engineers,” he said. “Everyone here cares deeply about what they build. We’re honest with each other, we experiment constantly, and we hold ourselves to a high technical standard.”
That culture, he added, is what allows Macrocosmos to thrive even in a fast-moving, competitive ecosystem.
IOTA and the Future of Distributed AI Training
The conversation naturally moved to IOTA, Macrocosmos’s flagship Subnet 9 project. Shifa asked Brian to break it down for a general audience.
He explained IOTA as a distributed system for training large neural networks, a way to split huge models into smaller parts that can be trained in parallel across different machines.
He compared it to a swarm of bees, each doing a small piece of the work. “If you can coordinate that swarm,” he said, “you unlock massive collective power without needing centralized data centers.”
For Shifa, the idea of AI training happening on personal computers instead of expensive clusters felt transformative. “It’s like bringing machine learning to the edge,” she observed and Brian agreed, calling it “one of the most exciting frontiers of decentralized computing.”
Research as a Core Value
Shifa also highlighted Brian’s Open Bittensor Research project on GitHub, an open repository for sharing academic and technical research from across the ecosystem.
Brian explained that it was born out of necessity. “There’s incredible research happening all over Bittensor,” he said, “but no single place to find it. I wanted to change that.”
He noted that Macrocosmos was founded as a research group and remains deeply committed to scientific inquiry. “Good research builds good products,” he said simply. “They feed each other.”
Bitstarter and the Next Wave of Builders
When the topic turned to Bitstarter, Brian’s energy shifted into something even more personal. Developed by Macrocosmos and led by Chris Zacharia, Bitstarter aims to become the first crowdfunding platform for Bittensor subnets.
Shifa asked him to explain how it works, and Brian described it as a system where new subnet teams can raise $TAO directly from the community, with backers earning alpha ownership in return. “It’s like a Kickstarter for Bittensor,” he said. “But it’s not just about money. It’s about mentorship, transparency, and giving new builders a fair start.”
The platform’s first project, Loosh AI, focuses on modeling human consciousness and robotics. This is a bold and unconventional choice that, according to Brian, embodies what Bitstarter is all about. “They believed in us early,” he said. “It felt right to give them the first spotlight.”
On the Halving and Staying Grounded
Shifa then asked about the upcoming $TAO halving, a topic of growing attention in the community. Brian’s response was measured. “No one knows what will happen,” he admitted. “The market could swing either way. My advice to new teams is simple: don’t obsess over numbers. Focus on building real value.”
He also reflected on how quickly the ecosystem evolves. “One day the emission structure changes, the next you’re recalibrating your entire model. It’s not easy, but that’s the beauty of building here. You can’t get too comfortable.”
The State of Bittensor
As the conversation wound down, Shifa invited Brian to reflect on Bittensor’s current state. He described it as “brilliant but unpredictable”, a network that constantly forces its participants to grow. “Bittensor makes you better,” he said. “It challenges how you think, how you design, and how you understand incentives. It’s complex, frustrating, but deeply beautiful.”
Shifa agreed, noting how every developer she’s spoken to expresses the same mix of challenge and devotion. “It’s rare,” she said, “to see technology that shapes people as much as they shape it.”
Looking Forward
In closing, Brian shared what keeps him motivated. “This ecosystem is filled with intelligent, passionate people. It’s not easy to build in, but it’s the right place to build. Every subnet, every experiment, is pushing the frontier of what’s possible.”
Shifa thanked him for the insight and candor, closing the discussion with a note that felt both hopeful and grounded: “Conversations like this remind us that Bittensor isn’t just code or economics; it’s people, ideas, and relentless curiosity coming together to build something greater.”
Be the first to comment